JP5143620B2 - Audition content distribution system and terminal device - Google Patents

Description

  The present invention relates to a trial content distribution system in which a mobile terminal device receives a plurality of types of content data for trial listening from a content distribution server and reproduces any of the received content data for trial listening.

  2. Description of the Related Art Conventionally, as a general music distribution service, a service that allows a user to listen to a chorus section of a music to be purchased is known. In this service, when a user selects a piece of music that he / she wants to audition, a sound source for trial listening of the selected music is reproduced. And if the user likes the song, it can be purchased.

  Patent Document 1 discloses a system (reproduction control method) that distributes contents for trial listening after adding various conditions such as a trial listening period. Patent Document 2 discloses a technology for distributing content from a certain content distribution device to a portable terminal by short-range wireless communication in a public place. This technology is characterized by providing guidance to portable terminals within the distribution device area. Patent Document 3 discloses a trial listening system for rental CD shop members. In this system, a member presents a member card and rents a portable terminal for listening to music. A trial listening history is recorded on the terminal. As a result, the system administrator can efficiently obtain the audition history of the member user.

  Further, Patent Document 4 discloses a system configured to set a valid period so that a trial listening portable terminal can be taken out of the store. Patent Document 5 discloses a technique that allows a trial listening player to audition a preview content at any time, and enables the purchase of the content at the timing of purchase. Patent Document 6 discloses a technology that allows music content (for sample listening) to be received and reproduced by a mobile terminal and purchased if it is liked.

In addition, as an existing service, a service that distributes a “program” produced by a content distribution side to mobile terminals of many users is known. This service can be regarded as a service related to "music preview" in that a mechanism for purchasing music flowing in the program is implemented by accessing information embedded in the program. . Furthermore, the existing digital recorder has a function of recording a large amount of television programs corresponding to keywords (genre, names of performers, etc.) designated by the user and presenting them to the user. With this function, it is possible to record a television program that the user wants to watch without specifying the program to be recorded each time.
JP 2001-282258 A JP 2003-022079 A JP 2006-040100 A JP 2006-040467 A JP 2006-079179 A JP 2001-028786 A JP 2003-316818 A K. Hoashi et al: Personalization of user profiles for content-based musical retrieval based on relevance fedback, Proc of ACM Multimedia 2003.

  However, the music preview function on the existing music distribution service requires the user to select the content he / she wants to audition. This is inconvenient and difficult to discover new music. In addition, the techniques described in Patent Document 1 to Patent Document 4 are inconvenient because they cannot be auditioned unless they go to a location (such as a rental DVD shop) where a terminal for audition is placed. In addition, in the technologies described in Patent Document 5 and Patent Document 6, since a mobile terminal is used, it is possible to audition at any time. However, since the user has to select (search) content to be auditioned, it is inefficient. It is.

  Further, in the case of a program distribution type service, since the music that the user wants to listen to is not necessarily distributed, there is little convenience for the user. Further, from the viewpoint of the content provider, it is necessary to produce a program (music composition) that has a high appeal to the user, which increases the cost. In addition, it is difficult to produce and distribute a program that suits individual user preferences. Furthermore, in a method of storing a large amount of content in a recorder having a large-capacity medium and allowing the user to select the content that he / she wants to see, it is necessary for the user to search for the content he / she is interested in from among the large amount of stored content. It is inefficient for the user. Also, in principle, only content that satisfies the conditions specified by the user is accumulated, and there is also a problem that it is difficult for the user to discover unknown content.

  The present invention has been made in view of such circumstances, and an object of the present invention is to provide a content distribution system for trial listening and a terminal device that are highly convenient for the user and can efficiently distribute the content for trial listening. And

  (1) In order to achieve the above object, the present invention takes the following measures. That is, the content distribution system for trial listening of the present invention includes a terminal device and a content distribution server, the terminal device receives a plurality of types of content data for trial listening from the content distribution server, and any of the received content for listening A content distribution system for trial listening that reproduces data, wherein the terminal device stores a content database that stores the received multiple types of content data for trial listening, a usage history information recording unit that records user usage history information, Based on the recorded user usage history information, a user preference estimation unit that estimates user preferences and generates preference information, the generated preference information, and each audition stored in the content database Calculating the degree of similarity with the content data, and the content database Et al, is characterized by comprising a contents extraction unit the calculated similarity is to extract high audition content data, content data reproduction unit for reproducing the audition content data the extracted, the.

  Thus, based on the user's usage history information, the user's preference is estimated and the preference information is generated, and the similarity between the generated preference information and each audition content data stored in the content database Since the content data for trial listening having a high degree of similarity is extracted from the content database, it is possible to efficiently deliver the content for trial listening that suits each user's preference. In addition, since the content data for audition can be reproduced without the user himself / herself setting conditions, the efficiency for the user is improved. In addition, since the user can present unknown content for trial listening, it is possible to create an opportunity to purchase new content that has been difficult with the prior art. Furthermore, by applying a mobile phone as a terminal device, there is no place restriction, and convenience for the user can be improved.

  (2) Moreover, in the content distribution system for trial listening of this invention, the said terminal device is triggered by selection operation by a user, either the content data for trial listening extracted by the said content extraction part, or the said content data reproduction | regeneration part A purchase content selection unit that generates purchase content information indicating that the content corresponding to the trial content data being reproduced or content corresponding to the trial content data reproduced by the content data reproduction unit is purchased, and the purchase content information And a purchase content information transmission unit for transmission to the content distribution server.

  In this way, a purchase indicating that one of the extracted content data for trial listening, or content corresponding to the content data for trial listening or the content data for trial listening that has been played is purchased, triggered by the selection operation by the user Since the content information is generated and the purchased content information is transmitted to the content distribution server, when the content data for trial listening is auditioned, it can be linked to the purchase.

  (3) Further, in the content distribution system for trial listening according to the present invention, the usage history information of the user includes at least one of a content purchase history, a content reproduction history, or a function registration history regarding content reproduction settings. It is characterized by.

  As described above, the user usage history information includes at least one of the content purchase history, the content playback history, or the function registration history related to the content playback setting. Therefore, when the user performs an operation related to the content on the terminal device. The history is recorded and can be used for selection of content data for trial listening. As a result, it is possible to efficiently distribute content for trial listening that suits individual user preferences. In addition, since the content data for audition can be reproduced without the user himself / herself setting conditions, the efficiency for the user is improved. In addition, since the user can present unknown content for trial listening, it is possible to create an opportunity to purchase new content that has been difficult with the prior art.

  (4) In the content distribution system for trial listening according to the present invention, the content data for trial listening is music data for trial listening, and the content extraction unit is stored in the generated preference information and the content database. The degree of similarity with each of the listening music data is calculated based on meta information added to the listening music data, an acoustic feature amount extracted from the music data, or a combination thereof.

  With this configuration, it is possible to extract music data according to the user's preference. As a result, it is possible to play the music data for trial listening that suits the taste of each user. In addition, since the music data for trial listening can be reproduced without the user himself / herself setting conditions, the efficiency for the user is improved. In addition, since the user can present unknown music data for trial listening, it is possible to create an opportunity to purchase new content that has been difficult with the prior art.

  (5) Also, in the content distribution system for trial listening of the present invention, the content distribution server classifies the stored content data for trial listening into a plurality of categories, a storage unit storing a plurality of types of content data for trial listening A content classifying unit, and a content transmitting unit that transmits the content data for trial listening classified into the plurality of categories to the terminal device, and the content database is transmitted from the content distribution server and is divided into a plurality of categories. The classified content data for trial listening is stored, and the content extraction unit extracts content data from any one of categories specified by a user.

  Thus, since the content data for trial listening stored in the content distribution server is classified into a plurality of categories, it is possible to efficiently extract content data according to the user's preference in the terminal device. .

  (6) The terminal device according to the present invention is a terminal device that receives a plurality of types of content data for trial listening from a content distribution server and reproduces any one of the received content data for trial listening. A content database for storing the content data for trial listening, a usage history information recording unit for recording the usage history information of the user, and the preference information by estimating the user's preference based on the recorded usage history information of the user The user preference estimation unit for generating the content, the generated preference information, and the content data for trial listening stored in the content database are calculated, and the calculated similarity is high from the content database. A content extraction unit for extracting content data for trial listening, and the extracted content data for trial listening It is characterized by and a content data reproducing unit for raw.

  Thus, based on the user's usage history information, the user's preference is estimated and the preference information is generated, and the similarity between the generated preference information and each audition content data stored in the content database Since the content data for trial listening having a high degree of similarity is extracted from the content database, it is possible to efficiently deliver the content for trial listening that suits each user's preference. In addition, since the content data for audition can be reproduced without the user himself / herself setting conditions, the efficiency for the user is improved. In addition, since the user can present unknown content for trial listening, it is possible to create an opportunity to purchase new content that has been difficult with the prior art. Furthermore, by applying a mobile phone as a terminal device, there is no place restriction, and convenience for the user can be improved.

  According to the present invention, based on user usage history information, user preference is estimated and preference information is generated, and the generated preference information and each audition content data stored in the content database are Since the similarity is calculated and the content data for trial listening having a high calculated similarity is extracted from the content database, it is possible to efficiently distribute the content for trial listening that suits the taste of each user. In addition, since the content data for audition can be reproduced without the user himself / herself setting conditions, the efficiency for the user is improved. In addition, since the user can present unknown content for trial listening, it is possible to create an opportunity to purchase new content that has been difficult with the prior art. Furthermore, by applying a mobile phone as a terminal device, there is no place restriction, and convenience for the user can be improved.

  In the embodiment of the present invention, a music content group for trial listening (such as a large number of music file groups for trial listening) is broadcasted to a user's mobile terminal device using, for example, BCMCS. The mobile terminal device preferentially presents and reproduces the content for trial listening that matches the user's preference on the application. If the user likes the music content that he / she listened to, he / she purchases it through a payment process. The information of “user preference” used when the application for trial listening in the mobile terminal device selects music to be presented to the user is estimated from, for example, the user's music playback history / purchase history.

  FIG. 1 is a diagram showing a schematic configuration of a content distribution system according to the present embodiment. The content distribution system 1 includes a content distribution server 10 that distributes content for trial listening and content purchased by a user, and a mobile terminal device 20 as a mobile phone used by the user. The content distribution server 10 and the mobile terminal device 20 can transmit and receive information via the network N.

  The content distribution server 10 has the following three modules. The trial listening content distribution module 11 distributes content for trial listening to a user who is a target of a service provided by the content distribution system according to the present embodiment. In the present embodiment, for example, a format in which the same audition content is distributed to a plurality of mobile terminal devices 20 by BCMCS or the like is assumed. The content purchase settlement module 12 receives the purchase content information selected by the user from the purchase content selection module 27 in the mobile terminal device 20, and performs settlement processing necessary for content purchase. In addition, the purchased content transmission module 13 transmits the content for which the payment processing has been completed in the content purchase settlement module 12 to the mobile terminal device 20 of the purchased user.

  On the other hand, the mobile terminal device 20 has the following seven modules. The music usage history storage module 21 stores music usage history information in the user's mobile terminal device 20. This “music usage history” includes music content purchase history, playback history, function registration history (eg, registration as ringtone), and the like. The user preference estimation module 22 estimates the user's music preference from the history information accumulated by the music use history accumulation module 21. The information generated by the user preference estimation module 22 is used in a process of selecting music that matches the user's preference from the trial listening content. A detailed processing method will be described later.

  The trial listening content receiving module 23 receives the trial listening content distributed from the trial listening content distribution module 11 of the content distribution server 10, and receives the trial listening content (sound source file + meta information) as the trial listening content of the mobile terminal device 20. Store in database 24. The audition content database 24 stores the audition content data received by the audition content receiving module 23.

  The user trial content selection module 25 compares the information of the content data for trial listening stored in the content database for trial listening 24 with the user preference information generated and output by the user preference estimation module 22, so that the user preference is selected. Select suitable content data for auditioning. A detailed processing method of the user trial listening content selection module 25 will be described later. The trial listening content reproduction module 26 reproduces the trial content data selected by the user trial content selection module 25. The content data for trial listening is reproduced as music from an audio output device such as the speaker 26a provided in the mobile terminal device 20, and the content of the content for trial listening being reproduced is displayed on the display (not shown) of the mobile terminal device 20. Processing for displaying information (for example, content title, performance artist, information for purchase, etc.) is performed.

  The purchase content selection module 27 designates content to be purchased by the user from among the content reproduced by the trial content reproduction module 26, triggered by an operation by the user. Note that these processes may be implemented as a function of the trial listening content playback module 26 described above. The purchase content information transmission module 28 transmits the purchase content information obtained from the purchase content selection module 27 to the content distribution server 10. Specifically, a process of transmitting to the content purchase settlement module 12 of the content distribution server 10 via the network N is performed.

  Next, the operation of the content distribution system according to this embodiment configured as described above will be described. The content distribution system according to the present embodiment extracts music information suitable for the user's musical preference from the content distribution server 10 from a large amount of content for trial listening distributed to the user's mobile terminal device 20, and The pre-listening content data extracted by the music pre-listening application on the terminal device 20 is preferentially reproduced to give a purchase opportunity. Specifically, the music use history is recorded in the user's portable terminal device 20, the user's music preference is estimated from the music use history, and the content data for the trial listening distributed based on the user's music preference is included. Content data that matches the user's preference is extracted.

(User preference estimation method from music usage history)
The “music use history” in the present embodiment includes, for example, the following information.
-Music playback history (number of times played, time, etc.) on the mobile terminal device 20
-Music content usage status on mobile terminal device 20 (registered as ringtone, etc.)
From the music use history information described above, it is possible to know information on music that the mobile terminal user likes. For example, music that is frequently played on the mobile terminal device 20 is presumed to be preferable music for the user. The same estimation can be made for music registered as a ring tone on the mobile terminal. Based on these pieces of information, user preference information is extracted by the following two methods.

(1) User preference extraction based on music meta information included in the reproduction history The music meta information is information indicating a performance artist, a genre, a release date, etc. of the music. In this method, for example, when a user has a large number of reproduction histories of a specific genre such as “lock”, “lock” can be extracted as a favorite genre of the user. In addition to simply extracting one piece of meta information (most frequent meta information in the playback history) as the user's preference information, the relative frequency in the playback history (for example, in the entire playback history) It is also possible to apply a method of expressing user preferences based on the appearance rate of each meta information.
(2) User preference extraction based on acoustic features of music included in playback history In the above method, for example, music corresponding to meta information extracted as information representing the user's music preference is included in the distributed listening content. If it is not included, it is difficult to extract music that suits the user's preference. Therefore, by applying a method for extracting the user's preference based on the acoustic features of the music included in the playback history, it is possible to extract the music that matches the user's preference without depending on the meta information of the sample music. Can do.

  One method for realizing the method (2) is the method proposed in Patent Document 7 and Non-Patent Document 1. Each of these documents proposes a method of vectorizing user preferences based on the acoustic features of music that the user likes. Specifically, based on the MFCC extracted from the music preferred by the user, after vectorizing each music by the tree vector quantization method, the vector representing the user preference is added by adding the vector of the music preferred by the user Is calculated.

(Music extraction method from sample content)
Here, the information included in the content for trial listening distributed from the content distribution server 10 (consisting of a large number of music pieces) includes the following.
・ Music sound source file ・ Music meta information (title, performance artist, genre)
-Music feature vector-Music purchase information

  By collating the above information and the user preference information extracted by the above-described method, music that can be auditioned by the user is extracted. For example, when a method for extracting user preferences using meta information of music is applied, by comparing the meta information included in the user preferences with the meta information of music included in the preview content, Sample content can be extracted. Specifically, when the user likes “rock”, a method of extracting a music piece whose genre is “rock” from the trial listening content is extracted. Also, as described above, when extracting user preference information based on the relative frequency in the reproduction history, the music for trial listening is selected from the content data for trial listening according to the frequency in the user preference information. The method of extracting can be applied.

On the other hand, when the user preference information is extracted based on the acoustic features of the music, the similarity between the user preference vector and the feature vector of the music in the preview content is calculated, and the music having a high similarity is extracted. Applicable methods are applicable. In particular,
The user preference vector is U = (u_1, u_2, ..., u_n),
Let the music acoustic vector be S = (s_1, s_2, ..., s_n),
Then, the similarity Sim (U, S) between the user preference U and the music S can be calculated by the following mathematical formula (1).

Sim (U, S) = (U · S) / | U || S |. . . . . (1)
However, U · S represents the inner product of the vectors U and S, and | U | represents the Euclidean length of the vector U.

  Here, for example, a method of calculating the similarity between the acoustic vectors of all the music pieces included in the content data for trial listening and the user's preference vector and extracting the top M pieces of music can be applied.

  In addition, a method of extracting sample music by integrating both meta information and acoustic features is also applicable. For example, a piece of music of a specific genre is extracted from the content data for trial listening as a subset, and a group of songs having a high (acoustic) similarity to user preference is extracted as a trial piece of music included in the subset. Techniques are listed.

  Next, a screen display example of the mobile terminal device 20 will be described. FIG. 2 is a diagram illustrating a screen display example in the mobile terminal device 20. In FIG. 2, the application screen is roughly divided into a “trial music display screen” and a “trial music playback screen”, and a sample music display screen on the left side and a test music playback screen on the right side in FIG. An example is shown. On the trial music display screen, a list of trial music extracted from the trial content is displayed by the method described above. Specifically, information on the music for trial listening (title, artist name, etc.) is displayed in the “trial music list” in the above figure. The user uses the operation keys of the mobile terminal to select a music piece to be auditioned. When the selected music is played, the screen shifts to the trial music playback screen. It is also possible to purchase the content data while being displayed on the preview music display screen.

  In addition, the trial music playback screen displays information on the trial music selected and played by the user on the trial music display screen, and has a function for purchasing the music being played by the user (or the music after playback). . Specifically, information (title, artist name, etc.) of the music being reproduced is displayed on the reproduced music information display section. When the user wants to purchase the music being played, the purchase process is performed by clicking the “playback music purchase button”. In this purchase process, information required for content purchase (ID of music to be purchased, user ID, etc.) is transmitted to the content purchase settlement module 12 of the content distribution server 10.

  In the above implementation example, the implementation mode in which the content trial listening application is installed on the mobile terminal device 20 is described. However, in the present invention, implementation on other terminals such as a set-top box (STB) is also possible. It is. In that case, the trial listening content is distributed via the Internet or the like. Further, in the case of an implementation having a trial listening content classification module, it is necessary to add a function for the user to select a content category that the user wants to audition for the content trial listening application. This function can be implemented in various forms. For example, a format in which a screen for allowing the user to select a category is inserted in front of the above "listening music display screen" is a suitable implementation form. Conceivable.

  FIG. 3 is a diagram showing a modification of the content distribution system according to the present embodiment. The content distribution system 100 is different from the content distribution system 1 shown in FIG. 1 in the configuration of the content distribution server 101. In the content distribution server 101 shown in FIG. 3, a “trial listening content classification module 102” is newly added. The audition content classification module 102 performs processing for classifying content distributed to each user into a plurality of categories. The classification result obtained by the trial listening content classification module 102 is passed to the trial listening content distribution module 11. The trial listening content distribution module 11 performs a process of distributing a plurality of classified content categories for trial listening to the user's portable terminal device 20 as trial content. The trial content reproduction module 26 on the mobile terminal device 20 has a function of selecting a category desired by the user from the distributed trial listening categories.

  The content classification method in the trial listening content classification module 102 is as follows. As a method of content classification processing for trial listening, a method of using each of meta information and acoustic features can be cited as in the above-described user preference extraction method. First, as a method for classifying content for trial listening based on meta information, a method of classifying content data for trial listening according to each meta information according to information (for example, genre information of music) given to the content data for trial listening can be applied. . Next, as a sample listening content classification method based on acoustic features, a clustering method using the above-described music feature vector can be considered. Specifically, the similarity between a predetermined cluster and the feature vector of the content to be classified is calculated, and the content is classified into a cluster having the highest similarity, or an existing method such as k-means clustering is used. Can be applied.

  As described above, according to the content distribution system according to the present embodiment, the present embodiment is compared with the format using the trial listening terminal proposed in Patent Documents 1 to 4 among the above-described conventional technologies. In this case, since it is assumed that a mobile terminal device is used, there is no restriction on the location, which is highly convenient for the user. Moreover, in the music trial listening / purchase format using the mobile terminal device proposed in the above-mentioned Patent Documents 5 and 6, etc., the user needs to select the music he / she wants to audition, but in this embodiment, This is improved. In addition, unlike conventional program delivery services, it is possible to efficiently deliver music for listening to music that suits individual user preferences, improving convenience for users and providing content providers. The cost on the side can also be reduced. Furthermore, since the user can efficiently reproduce the content to be reproduced without specifying the conditions by himself / herself, the efficiency for the user is greatly improved. Further, since it is possible to present unknown content for trial listening to the user, it is possible to provide an opportunity to purchase content that is difficult to be found by the conventional method.

It is a figure which shows schematic structure of the content delivery system which concerns on this embodiment. It is a figure which shows the example of a screen display in a portable terminal device. It is a figure which shows the modification of the content delivery system which concerns on this embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Content delivery system 10 Content delivery server 11 Content distribution module for trial listening 12 Content purchase settlement module 13 Purchased content transmission module 20 Mobile terminal device 21 Music use history storage module 22 User preference estimation module 23 Audition content reception module 24 Audition content database 25 User audition content selection module 26 Audition content playback module 26a Speaker 27 Purchased content selection module 28 Purchased content information transmission module 100 Content distribution system 101 Content distribution server 102 Audition content classification module

Claims (5)

A sample content distribution system, comprising a terminal device and a content distribution server, wherein the terminal device receives a plurality of types of sample content data from the content distribution server and reproduces any one of the received sample content data. ,
The content distribution server is
A storage unit for storing a plurality of types of content data for trial listening;
A content classification unit for classifying the stored content data for trial listening into a plurality of categories according to meta information or acoustic characteristics;
A content transmission unit that transmits the content data for trial listening classified into the plurality of categories to the terminal device,
The terminal device
A content database that is classified into a plurality of categories and stores content data for trial listening transmitted from the content distribution server ;
A usage history information recording unit for recording user usage history information;
Based on the recorded user usage history information, a user preference estimation unit that estimates user preferences and generates preference information;
The similarity between the generated preference information and each content data for trial listening stored in the content database is the inner product of the user preference vector based on the preference information and the music acoustic vector of the content data for trial listening. Calculated as a value obtained by dividing the vector by the product of the Euclidean lengths, and from the content database, the content data for trial listening having a high degree of similarity among the content data for trial listening of any one of the categories designated by the user A content extraction unit to extract;
A trial content distribution system comprising: a content data reproduction unit that reproduces the extracted content data for trial listening. The terminal device
Triggered by a user's selection operation, any of the content data for trial listening extracted by the content extraction unit, or content data for trial listening being played back by the content data playback unit or for trial playback played back by the content data playback unit A purchase content selection unit that generates purchase content information indicating that the content corresponding to the content data is to be purchased;
The trial content distribution system according to claim 1, further comprising: a purchase content information transmission unit configured to transmit the purchase content information to the content distribution server.   3. The sample listening history according to claim 1, wherein the usage history information of the user includes at least one of a content purchase history, a content playback history, or a function registration history related to content playback settings. Content distribution system. The content data for trial listening is music data for trial listening,
The content extraction unit extracts the similarity between the generated preference information and each piece of trial music data stored in the content database from the meta information added to the trial music data and the music data. The content distribution system for trial listening according to any one of claims 1 to 3, wherein the content is calculated based on the acoustic feature amount or a combination thereof. A terminal device that receives a plurality of types of content data for trial listening from a content distribution server and reproduces any of the received content data for trial listening,
A storage unit that stores a plurality of types of content data for trial listening, a content classification unit that classifies the stored content data for trial listening into a plurality of categories according to meta information or acoustic characteristics, and a classification into the plurality of categories A content database for storing trial content data transmitted from a content distribution server comprising a content transmission unit for transmitting the content data for trial listening ,
A usage history information recording unit for recording user usage history information;
Based on the recorded user usage history information, a user preference estimation unit that estimates user preferences and generates preference information;
The similarity between the generated preference information and each content data for trial listening stored in the content database is the inner product of the user preference vector based on the preference information and the music acoustic vector of the content data for trial listening. Calculated as a value obtained by dividing the vector by the product of the Euclidean lengths, and from the content database, the content data for trial listening having a high degree of similarity among the content data for trial listening of any one of the categories designated by the user A content extraction unit to extract;
A terminal device, comprising: a content data reproducing unit that reproduces the extracted content data for trial listening.

Images